CAREER FEATURE | 04 December 2020 | Correction 09 December 2020

How to write a superb literature review

Nature speaks to old hands and first timers about the work they did to make their reviews sing.

By Andy Tay

Credit: Getty

Literature reviews are important resources for scientists. They provide historical context for a
field while offering opinions on its future trajectory. Creating them can provide inspiration for
one’s own research, as well as some practice in writing. But few scientists are trained in how to
write a review — or in what constitutes an excellent one. Even picking the appropriate software
to use can be an involved decision (see ‘Tools and techniques’). So Nature asked editors and
working scientists with well-cited reviews for their tips.

WENTING ZHAO: Be focused and avoid jargon

Assistant professor of chemical and biomedical engineering, Nanyang Technological University,

Singapore.

When I was a research student, review writing improved my understanding of the history of my
field. I also learnt about unmet challenges in the field that triggered ideas.
For example, while writing my first review1 as a PhD student, I was frustrated by how poorly we
understood how cells actively sense, interact with and adapt to nanoparticles used in drug
delivery. This experience motivated me to study how the surface properties of nanoparticles can
be modified to enhance biological sensing. When I transitioned to my postdoctoral research, this
question led me to discover the role of cell-membrane curvature, which led to publications and
my current research focus. I wouldn’t have started in this area without writing that review.

A common problem for students writing their first reviews is being overly ambitious. When I
wrote mine, I imagined producing a comprehensive summary of every single type of
nanomaterial used in biological applications. It ended up becoming a colossal piece of work, with
too many papers discussed and without a clear way to categorize them. We published the work in
the end, but decided to limit the discussion strictly to nanoparticles for biological sensing, rather
than covering how different nanomaterials are used in biology.

My advice to students is to accept that a review is unlike a textbook: it should offer a more
focused discussion, and it’s OK to skip some topics so that you do not distract your readers.
Students should also consider editorial deadlines, especially for invited reviews: make sure that
the review’s scope is not so extensive that it delays the writing.

A good review should also avoid jargon and explain the basic concepts for someone who is new
to the field. Although I trained as an engineer, I’m interested in biology, and my research is about
developing nanomaterials to manipulate proteins at the cell membrane and how this can affect
ageing and cancer. As an ‘outsider’, the reviews that I find most useful for these biological topics
are those that speak to me in accessible scientific language.

Bozhi Tian likes to get a variety of perspectives into a review. Credit: Aleksander Prominski
BOZHI TIAN: Have a process and develop your style

Associate professor of chemistry, University of Chicago, Illinois.

In my lab, we start by asking: what is the purpose of this review? My reasons for writing one can
include the chance to contribute insights to the scientific community and identify opportunities
for my research. I also see review writing as a way to train early-career researchers in soft skills
such as project management and leadership. This is especially true for lead authors, because they
will learn to work with their co-authors to integrate the various sections into a piece with smooth
transitions and no overlaps.

After we have identified the need and purpose of a review article, I will form a team from the
researchers in my lab. I try to include students with different areas of expertise, because it is
useful to get a variety of perspectives. For example, in the review ‘An atlas of nano-enabled
neural interfaces’2, we had authors with backgrounds in biophysics, neuroengineering,
neurobiology and materials sciences focusing on different sections of the review.

After this, I will discuss an outline with my team. We go through multiple iterations to make sure
that we have scanned the literature sufficiently and do not repeat discussions that have appeared
in other reviews. It is also important that the outline is not decided by me alone: students often
have fresh ideas that they can bring to the table. Once this is done, we proceed with the writing.

I often remind my students to imagine themselves as ‘artists of science’ and encourage them to
develop how they write and present information. Adding more words isn’t always the best way:
for example, I enjoy using tables to summarize research progress and suggest future research
trajectories. I’ve also considered including short videos in our review papers to highlight key
aspects of the work. I think this can increase readership and accessibility because these videos can
be easily shared on social-media platforms.

ANKITA ANIRBAN: Timeliness and figures make a huge difference

Editor, Nature Reviews Physics.

One of my roles as a journal editor is to evaluate proposals for reviews. The best proposals are
timely and clearly explain why readers should pay attention to the proposed topic.

It is not enough for a review to be a summary of the latest growth in the literature: the most
interesting reviews instead provide a discussion about disagreements in the field.

Scientists often centre the story of their primary research papers around their figures — but when
it comes to reviews, figures often take a secondary role. In my opinion, review figures are more
important than most people think. One of my favourite review-style articles3 presents a plot
bringing together data from multiple research papers (many of which directly contradict each
other). This is then used to identify broad trends and suggest underlying mechanisms that could
explain all of the different conclusions.

An important role of a review article is to introduce researchers to a field. For this, schematic
figures can be useful to illustrate the science being discussed, in much the same way as the first
slide of a talk should. That is why, at Nature Reviews, we have in-house illustrators to assist
authors. However, simplicity is key, and even without support from professional illustrators,
researchers can still make use of many free drawing tools to enhance the value of their review
figures.

Yoojin Choi recommends that researchers be open to critiques when writing reviews. Credit: Yoojin Choi

YOOJIN CHOI: Stay updated and be open to suggestions

Research assistant professor, Korea Advanced Institute of Science and Technology, Daejeon.

I started writing the review ‘Biosynthesis of inorganic nanomaterials using microbial cells and
bacteriophages’4 as a PhD student in 2018. It took me one year to write the first draft because I
was working on the review alongside my PhD research and mostly on my own, with support from
my adviser. It took a further year to complete the processes of peer review, revision and
publication. During this time, many new papers and even competing reviews were published. To
provide the most up-to-date and original review, I had to stay abreast of the literature. In my case,
I made use of Google Scholar, which I set to send me daily updates of relevant literature based on
key words.

Through my review-writing process, I also learnt to be more open to critiques to enhance the
value and increase the readership of my work. Initially, my review was focused only on using
microbial cells such as bacteria to produce nanomaterials, which was the subject of my PhD
research. Bacteria such as these are known as biofactories: that is, organisms that produce
biological material which can be modified to produce useful materials, such as magnetic
nanoparticles for drug-delivery purposes.

However, when the first peer-review report came back, all three reviewers suggested expanding
the review to cover another type of biofactory: bacteriophages. These are essentially viruses that
infect bacteria, and they can also produce nanomaterials.
The feedback eventually led me to include a discussion of the differences between the various
biofactories (bacteriophages, bacteria, fungi and microalgae) and their advantages and
disadvantages. This turned out to be a great addition because it made the review more
comprehensive.

Writing the review also led me to an idea about using nanomaterial-modified microorganisms to
produce chemicals, which I’m still researching now.

PAULA MARTIN-GONZALEZ: Make good use of technology

PhD student, University of Cambridge, UK.

Just before the coronavirus lockdown, my PhD adviser and I decided to write a literature review
discussing the integration of medical imaging with genomics to improve ovarian cancer
management.

As I was researching the review, I noticed a trend in which some papers were consistently being
cited by many other papers in the field. It was clear to me that those papers must be important,
but as a new member of the field of integrated cancer biology, it was difficult to immediately find
and read all of these ‘seminal papers’.

That was when I decided to code a small application to make my literature research more
efficient. Using my code, users can enter a query, such as ‘ovarian cancer, computer tomography,
radiomics’, and the application searches for all relevant literature archived in databases such as
PubMed that feature these key words.

The code then identifies the relevant papers and creates a citation graph of all the references cited
in the results of the search. The software highlights papers that have many citation relationships
with other papers in the search, and could therefore be called seminal papers.

My code has substantially improved how I organize papers and has informed me of key
publications and discoveries in my research field: something that would have taken more time
and experience in the field otherwise. After I shared my code on GitHub, I received feedback that
it can be daunting for researchers who are not used to coding. Consequently, I am hoping to build
a more user-friendly interface in a form of a web page, akin to PubMed or Google Scholar, where
users can simply input their queries to generate citation graphs.

Tools and techniques

Most reference managers on the market offer similar capabilities when it comes to providing a
Microsoft Word plug-in and producing different citation styles. But depending on your working
preferences, some might be more suitable than others.

Reference managers
Attribute EndNote Mendeley Zotero Paperpile

Cost A one-time cost of around Free version available Free version Low and comes
US$340 but comes with available with academic
discounts
Attribute EndNote Mendeley Zotero Paperpile

discounts for academics;

around $150 for students

Level of user Extensive user tutorials Extensive user tutorials Forum Forum
support available; dedicated help available; global network of discussions to discussions to
desk 5,000 volunteers to advise troubleshoot troubleshoot
users

Desktop version Available Available Available Unavailable

available for
offline use?

Document Up to 2 GB (free version) Up to 2 GB (free version) Up to 300 MB Storage linked to

storage on cloud (free version) Google Drive

Compatible with No No Yes Yes

Google Docs?

Supports No group working References can be shared No limit on the No limit on the
collaborative or edited by a maximum of number of users number of users
working? three other users (or more
in the paid-for version)

Here is a comparison of the more popular collaborative writing tools, but there are other options,
including Fidus Writer, Manuscript.io, Authorea and Stencila.

Collaborative writing tools

Attribute Manubot Overleaf Google Docs

Cost Free, open source $15–30 per month, Free, comes with a Google
comes with academic account
discounts

Writing language Type and write in Markdown* Type and format in Standard word processor
LaTex*

Can be used with No No Yes

a mobile device?

References Bibliographies are built using Citation styles can be Possible but requires additional
DOIs, circumventing reference imported from reference referencing tools in a plug-in,
managers managers such as Paperpile

*Markdown and LaTex are code-based formatting languages favoured by physicists,

mathematicians and computer scientists who code on a regular basis, and less popular in
other disciplines such as biology and chemistry.
doi: https://doi.org/10.1038/d41586-020-03422-x

Interviews have been edited for length and clarity.

UPDATES & CORRECTIONS

Correction 09 December 2020: An earlier version of the tables in this article included some
incorrect details about the programs Zotero, Endnote and Manubot. These have now been corrected.

References
1. Hsing, I.-M., Xu, Y. & Zhao, W. Electroanalysis 19, 755–768 (2007).

Article Google Scholar

2. Ledesma, H. A. et al. Nature Nanotechnol. 14, 645–657 (2019).

Article PubMed Google Scholar

3. Brahlek, M., Koirala, N., Bansal, N. & Oh, S. Solid State Commun. 215–216, 54–62
(2015).

Article Google Scholar

4. Choi, Y. & Lee, S. Y. Nature Rev. Chem. https://doi.org/10.1038/s41570-020-00221-w

(2020).

Article PubMed Google Scholar

Download references